A process analysis of computer-generated patient reminders delivered by nursing personnel

Purpose: To measure the extent to which health care reminders are offered by nursing personnel and accepted by patients, we performed a detailed process study of a computer-generated, patient-specific mammography reminder. Subjects: 25,971 consecutive visits by 2,814 women age 50 and older seen during an eight year period in an academic internal medicine practice. Methods: An implementation trial with historical controls was done in the Internal Medicine Clinic at the University of Nebraska Medical Center. We incorporated a computer reminder system into the nursing check-in process along with the standing orders for mammograms. The administration of the reminder, the patient response to the reminder, and subsequent follow-through with mammography were tracked during the intervention phase. Patients were considered "up-to-date" if a mammogram had been done within two years before the visit, or if one was obtained within 60 days after the visit. Results: The proportion of patients up-to-date increased from 48% during a four-year pre-intervention phase to 54% in the intervention phase. Process analysis showed that clinic personnel offered the reminder in 31% of eligible visits, 11% of reminded patients agreed, and 53% of patients who agreed had mammography done. One unsuspected carcinoma was diagnosed. Patient race, insurance coverage, and number of visits during the prior two years influenced whether patients agreed to mammography and whether they then had the test done. Conclusions: Despite an intensive intervention incorporated into the routine care process, mammography adherence still fell short of optimal. Process analysis disclosed multiple points of failure. The lack of time during nursing check-in and patient decisions to refuse mammography were the major reasons for failure.